Rich:
Well, I did a little quick research on the subject (which admittedly
has not crossed my radar screen before) and I must say that you are pretty
much right on the money. This issue contains a mini-lesson in our old friend,
AC circuit theory, so bear with me for a minute.I think that those like me,
who are justing getting this news, might like to see a little serious bit of
explanation.
Imagine placing a small inductance in series with the filament. For
definiteness, let the filament hot resistance be 4 ohms and let the
inductance have an X(sub)L of 1 ohms. The effect of the 1 ohm of X will be
to raise the total filament load by only a few percent (Z=4.1<17 degrees) and
the power factor will become .97 (vs 1.0) which will cause a small increase
in the dissipation of both primary and secondary. So far, nothing earth
shaking has happened. Supposedly, the transformer primary could be tweaked a
few turns to keep the load voltage precise.
Now suppose we get a dead-short across the filament (approx. to the cold
start), and the secondary will see a load of X=1ohms and a Z of pure +j1. The
result will be, neglecting the R's at first, a secondary current of 4 times
the normal running current. It is a purely reactive load (nearly) so there
won't be much real power dissipation, for what that's worth. But the current
is safely limited to 400%, and could probably run like that for a while if
needed.
So with a little basic circuit strategy, the addition of some L, the current
limit is achieved. The next step is tricky and I am not going to go into
detail here. I state without proof (it is proven in books on transformer
design) that one of the effects of leakage inductance in a transformer is the
appearance of parasitic inductances in series with primary and secondary -
exactly as conjured up above. In fact, by strategically designing the
transformer and core to be less than perfect, an appropriate amount of
effective series inductance can be synthesized.
Sometime I would like to see the detailed construction of that transformer,
because it is not a totally trivial job to achieve exactly the desired amount
of leakage. Apparantly, there are several applications in the world where
this is common (e.g. neon transformers, AC welders, and oil burner ignition
transformers.)
73
Eric von Valtier K8LV
|